On measuring grounding and generalizing grounding problems
Daniel Quigley, Eric Maynard
TL;DR
This work reframes the symbol grounding problem as an auditable profile across authenticity, preservation, faithfulness, robustness, and compositionality, parameterized by context and meaning type. It introduces a formal framework tying symbols to meanings through agent-internal mappings and an evaluation pipeline, enabling precise measurement of grounding across symbolic, referential, vectorial, and relational modes. Through analyses of model-theoretic semantics, large language models, and natural language grounding, the authors show that none achieve full etiological grounding, with LLMs occupying an intermediate position between fully symbolic and fully embodied grounding. The framework is designed to guide targeted improvements, promote transparency about model understanding, and provide a common language for philosophers, computer scientists, linguists, and mathematicians to diagnose and compare grounding profiles across tasks and architectures.
Abstract
The symbol grounding problem asks how tokens like cat can be about cats, as opposed to mere shapes manipulated in a calculus. We recast grounding from a binary judgment into an audit across desiderata, each indexed by an evaluation tuple (context, meaning type, threat model, reference distribution): authenticity (mechanisms reside inside the agent and, for strong claims, were acquired through learning or evolution); preservation (atomic meanings remain intact); faithfulness, both correlational (realized meanings match intended ones) and etiological (internal mechanisms causally contribute to success); robustness (graceful degradation under declared perturbations); compositionality (the whole is built systematically from the parts). We apply this framework to four grounding modes (symbolic; referential; vectorial; relational) and three case studies: model-theoretic semantics achieves exact composition but lacks etiological warrant; large language models show correlational fit and local robustness for linguistic tasks, yet lack selection-for-success on world tasks without grounded interaction; human language meets the desiderata under strong authenticity through evolutionary and developmental acquisition. By operationalizing a philosophical inquiry about representation, we equip philosophers of science, computer scientists, linguists, and mathematicians with a common language and technical framework for systematic investigation of grounding and meaning.